Reconnaissance by close remote sensing of vineyards and analysis of their spatial structure using intra-plot frequency analysis. Application to the monitoring of the effects of cultural practices

Abstract

Under mediterranean viticultural conditions, rainfall and pollutant runoff are to a large extent determined by soil surface features and field structure. Recognising these surface variables at the scale of one or more catchments is mandatory for the comprehension of flooding events, agricultural pollution, and water resource management. The approach adopted by this study is to update vineyard information over large areas on the basis of airborne remote sensing, providing high resolution data. At the end this should provide field data to spatially distributed hydrological models. Within this context, colour image mosaics have been analysed on the basis of a per-field Fourier transform. An automatic analysis provides land cover, training mode (wire-trained/goblet) and geometry of vine rows (spacing, orientation and cover). The accurate and reproducible data enrich the field database. The method seems (i) transposable to other regions and other discontinuous crops, and (ii) an encouraging perspective for the study of other hydrologically interesting parameters (agricultural practises, weed cover...).